Problem 2: Challenge facing law enforcement
As noted in Crazy Technologies, innovations are very sci-fi-esq these days. So why isn’t law enforcement more readily using these systems? While I am admittedly against a surveillance state and still believe that face recognition is extreme and totally against my civil liberties (I mean what if I have a bad makeup day?), there seem to be endless arguments against using technologies in law enforcement. The ‘bad guys’ aren’t being biased why should we? In regards to cyber-crime, it is incredibly difficult to trace and determine who the perpetrators are. The exact tools and techniques involved in a discussion regarding safeguarding of computers or how to trace computer intruders is far beyond my knowledge base, so I will leave those arguments to the more technically savvy.
Instead, I’ll turn my attention to an information technology which has been debated about for sometime. I am talking about data-mining again. Those against using data-mining says that it is an infringement upon civil liberties, those pro-DM say an argument means we are sacrificing security. I am still a bit on the fence. But, the privacy advocates make a good point, and saying this from someone who has a high respect for high research, why use something that is faulty?
Here seem to be the big complaints against data-mining:
1. Mass databasing of personal information – as exemplified by the failed ‘Total Information Awareness’ Act, it would require those in the US to register all viable information about themselves. Ok, that would be somewhat creepy, so the mass database of information can be avoided. There will probably be a way to cross search government and private databases in a few years anyway.
2. Lack of Theory – Privacy advocates argue that law enforcement and other governmental analysts that employ data-mining techniques to sift through mass amounts of information do not base their searches, which often use algorithims, in theory. This is a problem. Because what is driving the searches expect perhaps investigator bias or stereotypes? And how can we ensure these searches are valid?
3. False Positive High Return – This issue is actually perpetuated by problem #2. When searches are done and they return results which are ‘false positive,’ meaning that someone was flagged as ‘terrorist’ when really they are not, can have a lot of problems. Because there is no theory behind searches, false positive rates are extremely high when the current data-mining techniques are employed.
Thus, how to fix this problem? 1. Forget about the database. 2. Build Theory 3. Test theories to reduce false positives – refine theory.